Asian Cardiovascular and Thoracic Annals http://aan.sagepub.com/
Factors associated with atrial fibrillation in rheumatic mitral stenosis Leili Pourafkari, Samad Ghaffari, George R Bancroft, Arezou Tajlil and Nader D Nader Asian Cardiovascular and Thoracic Annals published online 2 April 2014 DOI: 10.1177/0218492314530134 The online version of this article can be found at: http://aan.sagepub.com/content/early/2014/04/02/0218492314530134
Published by: http://www.sagepublications.com
On behalf of:
The Asian Society for Cardiovascular Surgery
Additional services and information for Asian Cardiovascular and Thoracic Annals can be found at: Email Alerts: http://aan.sagepub.com/cgi/alerts Subscriptions: http://aan.sagepub.com/subscriptions Reprints: http://www.sagepub.com/journalsReprints.nav Permissions: http://www.sagepub.com/journalsPermissions.nav
>> OnlineFirst Version of Record - Apr 2, 2014 What is This?
Downloaded from aan.sagepub.com at UNIV OF WASHINGTON LIBRARY on August 22, 2014
XML Template (2014) [27.3.2014–8:43am] //blrnas3/cenpro/ApplicationFiles/Journals/SAGE/3B2/AANJ/Vol00000/140063/APPFile/SG-AANJ140063.3d
(AAN)
[1–7] [PREPRINTER stage]
Original Article
Factors associated with atrial fibrillation in rheumatic mitral stenosis
Asian Cardiovascular & Thoracic Annals 0(0) 1–7 ß The Author(s) 2014 Reprints and permissions: sagepub.co.uk/journalsPermissions.nav DOI: 10.1177/0218492314530134 aan.sagepub.com
Leili Pourafkari1, Samad Ghaffari1, George R Bancroft2, Arezou Tajlil1 and Nader D Nader2
Abstract Background: Atrial fibrillation is a complication of mitral valve stenosis that causes several adverse neurologic outcomes. Our objective was to establish a mathematical model to predict the risk of atrial fibrillation in patients with mitral stenosis. Methods: Of 819 patients with mitral stenosis who were screened, 603 were enrolled in the study and grouped according to whether they were in sinus rhythm or atrial fibrillation. Demographic, echocardiographic, and hemodynamic data were recorded. Logistic regression models were constructed to identify the relative risks for each contributing factor and calculate the probability of developing atrial fibrillation. Receiver operating characteristic curves were plotted. Results: Two hundred (33%) patients had atrial fibrillation; this group was older, in a higher functional class, more likely to have suffered previous thromboembolic events, and had significantly larger left atrial diameters, lower ejection fractions, and lower left atrial appendage emptying flow velocity. The factors independently associated with atrial fibrillation were left atrial strain (odds ratio ¼ 7.53 [4.47–12.69], p < 0.001), right atrial pressure (odds ratio ¼ 1.09 [1.02–1.17], p ¼ 0.01), age (odds ratio ¼ 1.14 [1.05–1.25], p ¼ 0.002), and ejection fraction (odds ratio ¼ 0.92 [0.87–0.97], p ¼ 0.003). The area under the curve for the combined receiver operating characteristic for this model was 0.90 0.12. Conclusion: Age, right atrial pressure, ejection fraction, and left atrial strain can be used to construct a mathematical model to predict the development of atrial fibrillation in rheumatic mitral stenosis.
Keywords Atrial fibrillation, Logistic models, Mitral valve stenosis, Rheumatic heart disease, Risk factors
Introduction The most common etiology for mitral stenosis (MS) is rheumatic heart disease. Despite its decreased incidence in developed countries, rheumatic heart disease remains a major problem in the developing world. MS can be complicated by atrial fibrillation (AF) which affects approximately 40% of patients with this disorder. AF predisposes patients with MS to thromboembolic events.1 The occurrence of AF in MS is multifactorial in nature. Elevated levels of high-sensitivity C-reactive protein and N-terminal brain natriuretic peptide precursors are reported in affected patients.2,3 Studies have evaluated the influence of echocardiographic and catheterization findings in patients with AF, but the results have been inconsistent.4–6 The stagnant blood flow in the left atrial appendage (LAA) that occurs in the absence of forcible atrial contraction during
fibrillation initiates the clot-forming processes and leads to the development of arterial thromboembolic events in MS patients.7–10 The prognosis of MS patients with thromboembolic stroke is poor, and their quality of life and functional status are often more limited by neurologic deficits than by the symptoms of MS. Therefore, it is important to identify the factors that predispose MS patients to developing AF, which may help clinicians better manage susceptible patients. 1 Cardiovascular Research Center, Tabriz University of Medical Sciences, Tabriz, Iran 2 Department of Anesthesiology, University at Buffalo, Buffalo, NY, USA
Corresponding author: Nader D Nader, MD, PhD, FCCP, Department of Anesthesiology, Main Campus 252 Farber Hall, Buffalo, NY 14214, USA. Email: [email protected]
Downloaded from aan.sagepub.com at UNIV OF WASHINGTON LIBRARY on August 22, 2014
XML Template (2014) [27.3.2014–8:43am] //blrnas3/cenpro/ApplicationFiles/Journals/SAGE/3B2/AANJ/Vol00000/140063/APPFile/SG-AANJ140063.3d
(AAN)
[1–7] [PREPRINTER stage]
2
Asian Cardiovascular & Thoracic Annals 0(0)
In this regard, we retrospectively evaluated our prospectively maintained patient database to assess differences between patients with sinus rhythm (SR) and those with AF. The main objective of the study was to determine the predictive factors for AF and establish a mathematical prediction model that may be used for calculation of the clinical risk of AF in MS. We hypothesized that a combination of demographic and echocardiographic findings could be used to reliably predict the occurrence of AF.
Patients and methods We conducted a retrospective case-control study on patients with a confirmed diagnosis of rheumatic MS who were scheduled to undergo a percutaneous transvenous mitral commissurotomy (PTMC) procedure in Tabriz Heart Center between March 2002 and March 2012. The study was approved by the institutional review board at Tabriz University of Medical Sciences, and exempted from informed consent due to its retrospective design. However, data were treated carefully to maintain complete patient privacy. Of 819 patients who were screened, 603 were included in the study. Patients with a prior diagnoses of hypertension (n ¼ 135), hyperthyroidism (n ¼ 5), ischemic heart disease (n ¼ 19), aortic valve stenosis of more than mild severity, and aortic insufficiency or mitral regurgitation of more than moderate severity (n ¼ 57), were excluded. The patients were classified as in SR or AF based on their baseline electrocardiogram. AF was defined as an irregularly irregular heart rate without any detectable P waves along with fibrillation f waves on 12-lead electrocardiograms. Demographic, echocardiographic, and hemodynamic data were recorded. The studied demographic data included age, sex, diabetes mellitus, smoking, blood group, and previous thromboembolic events. Echocardiographic data were recorded from transesophageal echocardiography before PTMC or surgery. Mitral valve area, left atrial diameter, left atrial appendage emptying velocity, left atrial appendage filling velocity, left ventricular (LV) ejection fraction, LV end-diastolic diameter, LV end-systolic diameter, and tricuspid valve regurgitation gradient were the studied echocardiographic variables. Additionally, left atrial strain was measured by the speckle tracking technique, as described by Shaikh and colleagues.11 The Vivid 7 (GE Healthcare, Waukesha, WI, USA) machine is equipped with software that automates the rest of the measurements after Doppler imaging. Tissue speckles were tracked rather than left atrial endothelium, and the mean value of 3 measurements are presented. Hemodynamic data obtained before and after the PTMC procedure included left atrial pressure, right atrial pressure, right ventricular systolic pressure, left
atrial strain, systolic and diastolic pulmonary artery pressures, and LV systolic and diastolic pressures. The presence of coronary artery disease (in patients who underwent coronary angiography) was also recorded. The baseline electrocardiograms were evaluated for signs of right ventricular hypertrophy which was determined by a qR pattern or R/S >1 in lead V1. Data were collected and recorded in Microsoft Excel workbook and exported to SPSS version 18.0 (SPSS, Inc., Chicago, IL, USA) for data analyses. AF was the main outcome binary variable. Various categorical and numerical variables from echocardiography and hemodynamic studies were used as independent variables to calculate the risk of AF. Fisher’s exact test with chisquare analysis were performed to compare the frequencies of categorical variables, and Student’s t test was used to compare numerical variables between the 2 groups. Categorical variables are reported as number and percentage, and continuous variables are expressed as mean standard deviation. The variables that were significantly different between the two groups and had the least multi-colinearity were used in a multiple logistic regression model to detect predictive factors for AF. Receiver operating characteristic curves were plotted, and the area under the curve values were used to calculate the overall predictive value of the model and its sensitivity-specificity ratios. Area under the curve values >0.7 were considered to be reasonable predictive models. Confidence intervals and odds ratios are given for these factors after both univariate and multivariate analyses (if applicable); p values less than 0.05 were considered statistically significant.
Results Among 603 patients with MS, 200 were found to be in AF (33%). Patients with AF were significantly older than those in the SR group (Table 1). Patients with AF had more central nervous system thromboembolic events compared to those in the SR group. More patients in the AF group reported dyspnea with minimal exertion or at rest (New York Heart Association class III or IV). Left atrial diameters were larger in patients with AF (Table 2). Left atrial volume was calculated from the left atrial diameter obtained from 2 right-angle axes (apical 4-chamber and 2-chamber views). Similar to left atrial diameter, left atrial volume was greater in patients with AF. Both LAA filling and emptying flow velocities were significantly lower in patients with AF. Although LV end-systolic and end-diastolic diameters were similar in both groups, the LV ejection fraction was modestly higher in the SR group. There was no difference in mitral valve area or tricuspid valve regurgitation gradients between patients with AF and SR (Table 2).
Downloaded from aan.sagepub.com at UNIV OF WASHINGTON LIBRARY on August 22, 2014
XML Template (2014) [27.3.2014–8:43am] //blrnas3/cenpro/ApplicationFiles/Journals/SAGE/3B2/AANJ/Vol00000/140063/APPFile/SG-AANJ140063.3d
(AAN)
[1–7] [PREPRINTER stage]
Pourafkari et al.
3
Table 1. Demographic characteristics and comorbid conditions in 603 patients with rheumatic mitral stenosis. Variable
AF group (n ¼ 200)
SR group (n ¼ 403)
p value
Age (years) Sex (F/M) Body mass index (kgm2) Diabetes mellitus Preexisting CNS emboli Current smoker Functional class III/IV Right ventricular hypertrophy
52.2 11.1 148/52 25.1 4.3 5 (2.5%) 22 (11%) 18 (9%) 56.1% 56 (29.3%)
39.9 11.6 317/86 25.2 4.4 9 (2.2%) 8 (2%) 28 (6.9%) 37.6% 122 (31.4%)
Factors associated with atrial fibrillation in rheumatic mitral stenosis Leili Pourafkari, Samad Ghaffari, George R Bancroft, Arezou Tajlil and Nader D Nader Asian Cardiovascular and Thoracic Annals published online 2 April 2014 DOI: 10.1177/0218492314530134 The online version of this article can be found at: http://aan.sagepub.com/content/early/2014/04/02/0218492314530134
Published by: http://www.sagepublications.com
On behalf of:
The Asian Society for Cardiovascular Surgery
Additional services and information for Asian Cardiovascular and Thoracic Annals can be found at: Email Alerts: http://aan.sagepub.com/cgi/alerts Subscriptions: http://aan.sagepub.com/subscriptions Reprints: http://www.sagepub.com/journalsReprints.nav Permissions: http://www.sagepub.com/journalsPermissions.nav
>> OnlineFirst Version of Record - Apr 2, 2014 What is This?
Downloaded from aan.sagepub.com at UNIV OF WASHINGTON LIBRARY on August 22, 2014
XML Template (2014) [27.3.2014–8:43am] //blrnas3/cenpro/ApplicationFiles/Journals/SAGE/3B2/AANJ/Vol00000/140063/APPFile/SG-AANJ140063.3d
(AAN)
[1–7] [PREPRINTER stage]
Original Article
Factors associated with atrial fibrillation in rheumatic mitral stenosis
Asian Cardiovascular & Thoracic Annals 0(0) 1–7 ß The Author(s) 2014 Reprints and permissions: sagepub.co.uk/journalsPermissions.nav DOI: 10.1177/0218492314530134 aan.sagepub.com
Leili Pourafkari1, Samad Ghaffari1, George R Bancroft2, Arezou Tajlil1 and Nader D Nader2
Abstract Background: Atrial fibrillation is a complication of mitral valve stenosis that causes several adverse neurologic outcomes. Our objective was to establish a mathematical model to predict the risk of atrial fibrillation in patients with mitral stenosis. Methods: Of 819 patients with mitral stenosis who were screened, 603 were enrolled in the study and grouped according to whether they were in sinus rhythm or atrial fibrillation. Demographic, echocardiographic, and hemodynamic data were recorded. Logistic regression models were constructed to identify the relative risks for each contributing factor and calculate the probability of developing atrial fibrillation. Receiver operating characteristic curves were plotted. Results: Two hundred (33%) patients had atrial fibrillation; this group was older, in a higher functional class, more likely to have suffered previous thromboembolic events, and had significantly larger left atrial diameters, lower ejection fractions, and lower left atrial appendage emptying flow velocity. The factors independently associated with atrial fibrillation were left atrial strain (odds ratio ¼ 7.53 [4.47–12.69], p < 0.001), right atrial pressure (odds ratio ¼ 1.09 [1.02–1.17], p ¼ 0.01), age (odds ratio ¼ 1.14 [1.05–1.25], p ¼ 0.002), and ejection fraction (odds ratio ¼ 0.92 [0.87–0.97], p ¼ 0.003). The area under the curve for the combined receiver operating characteristic for this model was 0.90 0.12. Conclusion: Age, right atrial pressure, ejection fraction, and left atrial strain can be used to construct a mathematical model to predict the development of atrial fibrillation in rheumatic mitral stenosis.
Keywords Atrial fibrillation, Logistic models, Mitral valve stenosis, Rheumatic heart disease, Risk factors
Introduction The most common etiology for mitral stenosis (MS) is rheumatic heart disease. Despite its decreased incidence in developed countries, rheumatic heart disease remains a major problem in the developing world. MS can be complicated by atrial fibrillation (AF) which affects approximately 40% of patients with this disorder. AF predisposes patients with MS to thromboembolic events.1 The occurrence of AF in MS is multifactorial in nature. Elevated levels of high-sensitivity C-reactive protein and N-terminal brain natriuretic peptide precursors are reported in affected patients.2,3 Studies have evaluated the influence of echocardiographic and catheterization findings in patients with AF, but the results have been inconsistent.4–6 The stagnant blood flow in the left atrial appendage (LAA) that occurs in the absence of forcible atrial contraction during
fibrillation initiates the clot-forming processes and leads to the development of arterial thromboembolic events in MS patients.7–10 The prognosis of MS patients with thromboembolic stroke is poor, and their quality of life and functional status are often more limited by neurologic deficits than by the symptoms of MS. Therefore, it is important to identify the factors that predispose MS patients to developing AF, which may help clinicians better manage susceptible patients. 1 Cardiovascular Research Center, Tabriz University of Medical Sciences, Tabriz, Iran 2 Department of Anesthesiology, University at Buffalo, Buffalo, NY, USA
Corresponding author: Nader D Nader, MD, PhD, FCCP, Department of Anesthesiology, Main Campus 252 Farber Hall, Buffalo, NY 14214, USA. Email: [email protected]
Downloaded from aan.sagepub.com at UNIV OF WASHINGTON LIBRARY on August 22, 2014
XML Template (2014) [27.3.2014–8:43am] //blrnas3/cenpro/ApplicationFiles/Journals/SAGE/3B2/AANJ/Vol00000/140063/APPFile/SG-AANJ140063.3d
(AAN)
[1–7] [PREPRINTER stage]
2
Asian Cardiovascular & Thoracic Annals 0(0)
In this regard, we retrospectively evaluated our prospectively maintained patient database to assess differences between patients with sinus rhythm (SR) and those with AF. The main objective of the study was to determine the predictive factors for AF and establish a mathematical prediction model that may be used for calculation of the clinical risk of AF in MS. We hypothesized that a combination of demographic and echocardiographic findings could be used to reliably predict the occurrence of AF.
Patients and methods We conducted a retrospective case-control study on patients with a confirmed diagnosis of rheumatic MS who were scheduled to undergo a percutaneous transvenous mitral commissurotomy (PTMC) procedure in Tabriz Heart Center between March 2002 and March 2012. The study was approved by the institutional review board at Tabriz University of Medical Sciences, and exempted from informed consent due to its retrospective design. However, data were treated carefully to maintain complete patient privacy. Of 819 patients who were screened, 603 were included in the study. Patients with a prior diagnoses of hypertension (n ¼ 135), hyperthyroidism (n ¼ 5), ischemic heart disease (n ¼ 19), aortic valve stenosis of more than mild severity, and aortic insufficiency or mitral regurgitation of more than moderate severity (n ¼ 57), were excluded. The patients were classified as in SR or AF based on their baseline electrocardiogram. AF was defined as an irregularly irregular heart rate without any detectable P waves along with fibrillation f waves on 12-lead electrocardiograms. Demographic, echocardiographic, and hemodynamic data were recorded. The studied demographic data included age, sex, diabetes mellitus, smoking, blood group, and previous thromboembolic events. Echocardiographic data were recorded from transesophageal echocardiography before PTMC or surgery. Mitral valve area, left atrial diameter, left atrial appendage emptying velocity, left atrial appendage filling velocity, left ventricular (LV) ejection fraction, LV end-diastolic diameter, LV end-systolic diameter, and tricuspid valve regurgitation gradient were the studied echocardiographic variables. Additionally, left atrial strain was measured by the speckle tracking technique, as described by Shaikh and colleagues.11 The Vivid 7 (GE Healthcare, Waukesha, WI, USA) machine is equipped with software that automates the rest of the measurements after Doppler imaging. Tissue speckles were tracked rather than left atrial endothelium, and the mean value of 3 measurements are presented. Hemodynamic data obtained before and after the PTMC procedure included left atrial pressure, right atrial pressure, right ventricular systolic pressure, left
atrial strain, systolic and diastolic pulmonary artery pressures, and LV systolic and diastolic pressures. The presence of coronary artery disease (in patients who underwent coronary angiography) was also recorded. The baseline electrocardiograms were evaluated for signs of right ventricular hypertrophy which was determined by a qR pattern or R/S >1 in lead V1. Data were collected and recorded in Microsoft Excel workbook and exported to SPSS version 18.0 (SPSS, Inc., Chicago, IL, USA) for data analyses. AF was the main outcome binary variable. Various categorical and numerical variables from echocardiography and hemodynamic studies were used as independent variables to calculate the risk of AF. Fisher’s exact test with chisquare analysis were performed to compare the frequencies of categorical variables, and Student’s t test was used to compare numerical variables between the 2 groups. Categorical variables are reported as number and percentage, and continuous variables are expressed as mean standard deviation. The variables that were significantly different between the two groups and had the least multi-colinearity were used in a multiple logistic regression model to detect predictive factors for AF. Receiver operating characteristic curves were plotted, and the area under the curve values were used to calculate the overall predictive value of the model and its sensitivity-specificity ratios. Area under the curve values >0.7 were considered to be reasonable predictive models. Confidence intervals and odds ratios are given for these factors after both univariate and multivariate analyses (if applicable); p values less than 0.05 were considered statistically significant.
Results Among 603 patients with MS, 200 were found to be in AF (33%). Patients with AF were significantly older than those in the SR group (Table 1). Patients with AF had more central nervous system thromboembolic events compared to those in the SR group. More patients in the AF group reported dyspnea with minimal exertion or at rest (New York Heart Association class III or IV). Left atrial diameters were larger in patients with AF (Table 2). Left atrial volume was calculated from the left atrial diameter obtained from 2 right-angle axes (apical 4-chamber and 2-chamber views). Similar to left atrial diameter, left atrial volume was greater in patients with AF. Both LAA filling and emptying flow velocities were significantly lower in patients with AF. Although LV end-systolic and end-diastolic diameters were similar in both groups, the LV ejection fraction was modestly higher in the SR group. There was no difference in mitral valve area or tricuspid valve regurgitation gradients between patients with AF and SR (Table 2).
Downloaded from aan.sagepub.com at UNIV OF WASHINGTON LIBRARY on August 22, 2014
XML Template (2014) [27.3.2014–8:43am] //blrnas3/cenpro/ApplicationFiles/Journals/SAGE/3B2/AANJ/Vol00000/140063/APPFile/SG-AANJ140063.3d
(AAN)
[1–7] [PREPRINTER stage]
Pourafkari et al.
3
Table 1. Demographic characteristics and comorbid conditions in 603 patients with rheumatic mitral stenosis. Variable
AF group (n ¼ 200)
SR group (n ¼ 403)
p value
Age (years) Sex (F/M) Body mass index (kgm2) Diabetes mellitus Preexisting CNS emboli Current smoker Functional class III/IV Right ventricular hypertrophy
52.2 11.1 148/52 25.1 4.3 5 (2.5%) 22 (11%) 18 (9%) 56.1% 56 (29.3%)
39.9 11.6 317/86 25.2 4.4 9 (2.2%) 8 (2%) 28 (6.9%) 37.6% 122 (31.4%)
